DeepSig

DeepSig: as Clarke’s third law says, any sufficiently advanced technology is indistinguishable from magic. It’s hard to imagine just how incredibly complex a system you need to ensure that the 5 billion people on this planet who carry a phone around with them have access to a cellular network. Essentially, you carry a device in your pocket that lets you communicate instantly with two-thirds of the people on this planet. (And if the man gets you down, you can use the device to tell 47% of the planet on social media how oppressed you are.) One key component of this incredible global telecommunications system is signal processing.

Founded in 2016, Virginia startup DeepSig has taken in $7.7 million in funding from investors that include Lockheed Martin. The products that DeepSig has developed use machine learning to perform functions such as dealing with radio frequency (RF) interference which causes calls to drop, signals to be poor, and overall creates a bad user experience. A case study DeepSig published talks about how “Leveraging Artificial Intelligence (AI) at the edge opens new frontiers in Radio Frequency (RF) signal analysis.”

“At the edge” refers to doing things on the device rather than up in the cloud. So, DeepSig is using machine learning algorithms at the edge for signal processing. They’re doing what so many other companies are doing – transferring the complexity from hardware to software. They’re also not alone when it comes to transferring intelligence to the edge.